Recomputable On-Chain Generative Art

Diving deeper into

Erick Calderon, CEO of Art Blocks, on the evolution of NFT marketplaces

Interview
If you're generating an Art Blocks output in 10 years, it will look exactly the same on that higher resolution screen
Analyzed 5 sources

The core advantage is permanence of the artwork as a process, not just permanence of a picture. Art Blocks stores the recipe and the token specific seed, so the same token can be rendered again at whatever screen size exists later, instead of stretching a fixed JPEG made for an older display. That is what makes the piece feel more like software that keeps producing the same result than a static file frozen at one resolution.

  • On Art Blocks, the collector does not buy a pre rendered image. The artist uploads code, the token stores a hash, and the generator combines them into the final output. That is why the image can be recomputed consistently, rather than merely downloaded from storage.
  • This also explains the difference versus most NFT marketplaces. A typical NFT points to an already made media file. Art Blocks mints the artwork at purchase, then later marketplaces like OpenSea display either a rendered thumbnail or the live algorithmic output.
  • The model has moved closer to full preservation over time. In 2022, common libraries like p5.js could still sit outside the chain. By 2026, Art Blocks says about 90% of projects are fully on-chain including dependencies, which makes the long term rendering promise much stronger.

Going forward, the market value of on-chain generative art should increasingly center on works whose rendering stack can survive independently of any one company website. That favors platforms that preserve not only ownership records, but the full instructions needed to regenerate the art decades later, across new devices, screens, and interfaces.